An automatic optimization tool for semiconductor devices based on iterative device simulations is developed. The tool is used for optimization of different kinds of semiconductor devices using various performance measures. High performance optimization algorithms, both local and global, are used to achieve an efficient design in shortest possible time. In this paper the effects of different optimization algorithms, performance measures, and number of variables in the optimization are studied. Both the computational efficiency and the devices achieved with different performance measures are studied. We give a demonstration of the usefulness of this method in a comparison between different device topologies, which have been optimized for best performance. (C) 2004 Elsevier Ltd. All rights reserved.